Corporate Ownership Structure in Japan Recent Trends and Their Impact

Corporate Ownership Structure in Japan Recent Trends and Their Impact by Keisuke Nitta Financial Research Group nitta@nli-research.co.jp The corporate ownership structure in Japan has changed significantly since the late 1990s. Using a new shareholder classification, we analyze ownership trends and their impact on corporate management. Among other things, our results reconfirm that unwinding of cross-shareholdings and growing institutional ownership have helped to enhance management discipline at public companies. 1. Introduction The major corporations in Japan do not have a single majority shareholder with a controlling interest. Generally, firms have a long list of large shareholders, each holding approximately 3% to 5% of outstanding shares. This widely dispersed ownership structure, which is characterized by the absence of controlling shareholders, is also common at public companies in the U.S. and U.K. However, in many cases, large shareholders at these firms do not hold diversified investment portfolios. This is partly because aside from large financial institutions, most investors such as business firms and individuals have limited funds available for stock investment. According to modern portfolio theory, concentrated investment is inefficient. When investors concentrate rather than diversify investment, they must pay comparatively high prices for rather risky assets. This raises the question as to why some shareholders still prefer to concentrate investment in a particular firm. There are two hypotheses explaining this behavior. The first one emphasizes the shared benefits of control large shareholders seek benefits deriving from enhanced firm value when they oversee management, which are benefits that minority shareholders also coincidentally enjoy. The second one emphasizes the private benefits of control large shareholders seek benefits from the exploitation of corporate resources and other means when they exercise significant influence over the company s management. In this case, they enjoy benefits at the expense of minority shareholders. Both objectives can coexist in a shareholder, and it depends on the shareholder type as to which characteristic is stronger. On the other hand, institutional investors have ample resources to diversify investment and still become large shareholders. Since institutional investors seek to maximize investment returns for their clients, the above conflict of interest should not apply to them. Their main impact on corporate management derives from the monitoring stance they assume, and from their collective influence as measured by summing up the number of shares they hold. In Japan, shareholding is widespread at public companies, typically through cross-shareholding and ownership of listed subsidiaries. However, the incentive structure differs from that described above. Cross-shareholders are silent partners purposely created by the management, while parent firms exemplify the control structure of the corporate group. Thus to analyze the effect of ownership structure on corporate management, we need to classify NLI Research 1 2008.03.31

shareholder types based on their investment objective and incentive structure. From this new perspective, below we analyze recent trends in ownership structure and the implications for corporate management. 1 the aggregate shareholding ratio of institutional investors could become a key factor for disciplining the corporate management. As the ratio rises, so does the pressure on management to perform well. 2. Classification of Shareholder Types Unfortunately, no direct data is available to identify the investment objective or incentive structure of each shareholder. Thus we try to characterize shareholder types based of their observable attributions. For example, large individual shareholders whether they own shares directly or indirectly through an owned company presumably have both incentives mentioned above (shared benefits and private benefits of control). As for institutional investors, some of them actively monitor management and others do not. But they might reach the same conclusion that they should oppose the incumbent management if there is a suspicion that the management s decision-making is not efficient, because generally their investment objective is to earn adequate returns on investment. For this reason, Based on such considerations, we define shareholder types as shown in Exhibit 1. In general, since shareholders are required to hold a block of shares in order to exert influence over corporate management, we set the minimum threshold at 3%. 2 The classification method relies on shareholder names reported in the major shareholders database compiled by Toyo Keizai Inc. In ambiguous cases, we searched for further identifying information using the mass media, financial statement reports, and Internet websites. Due to data limitations, this identification should be regarded as a rough proxy for shareholder type. As mentioned earlier, there are some types of shareholders who do not have any influence on corporate management individually, but do have an impact if the total amount of shares represented by a particular shareholder type sums up to a considerable amount. Since we Exhibit 1 Classification of Shareholder Types Shareholder type Definition Cross-shareholder Financial institution * (excludes cross-shareholding) Listed firm (excludes cross-shareholding and financial institution) Foreign firm * Insider Director Domestic private firm * Foreign private firm and large individual * Stock ownership plan Govt. & public institution * Institutional investor Domestic Foreign Other minority shareholder Total cross-shareholding ratio. Banks, life insurers, and non-life insurers with at least 3% shareholding ratio; excludes crossshareholding. Publicly listed companies with at least 3% shareholding ratio; excludes cross-shareholding and financial institutions, but includes non-listed companies affiliated with listed holding companies. Foreign-registered business firms with at least 3% shareholding. Executives and auditors; includes directors' stock ownership plan. Domestic-registered private firms with at least 3% shareholding ratio. Foreign-registered private firms and individuals with at least 3% shareolding ratio. Stock ownership plans of employees and business partners. National and local government entities with at least 3% shareholding ratio. Pension trust, investment trust, life insurance special account. Foreigners excluding foreign-registered firms and large foreign individual shareholders. All other shareholders * For these shareholder types, the minimum shareholding threshold is set at 3%. NLI Research 2 2008.03.31

assume that each shareholder type shares some common interests, their aggregate shareholding ratio becomes a key factor. Aside from institutional investors as discussed earlier, this characteristic applies to cross-shareholders. 3 An important function of cross-shareholding is to fend off the actions of hostile shareholders. In doing so, the collective ratio of cross-shareholdings becomes more important than the relationship with a particular cross-shareholder. For corporate directors, however, the 3% threshold appears to be inappropriate. While the wealth effect of ownership provides a strong incentive, directors are less likely to become large shareholders as firm size increases because block ownership is financially more difficult to achieve. In addition, directors do not need large shareholdings to exert influence. Thus we believe shareholding ratios below 3% should be taken into account. Based on the above considerations, we set a 3% minimum threshold for those shareholders marked with an asterisk in Exhibit 1, and not for others. 4 Due to space limitations, a detailed explanation of the compilation method has been omitted. 3. Trends in Ownership Structure Based on the above shareholder type classification, we estimated the ownership structure at all publicly traded firms listed on the First Section of the Tokyo, Osaka, and Nagoya Stock Exchanges at each fiscal yearend from 1987 to 2006 (20 periods). Exhibit 2 shows the trend of the average shareholding ratio by each type of sample firm. Several important trends appear in the ownership structure of major firms in Japan. First, the shareholding ratio of institutional investors has surged dramatically in recent years. While the ratio rises over the entire period, the pace accelerates sharply in 2000, which is the year that institutional investors become the largest shareholder type. This trend suggests that the stock market is playing an increasingly active role in disciplining corporate management. The second trend is the decline of cross-shareholding. After stabilizing at 15% to 1996, it began steadily declining to around 8% in 2006. 5 Large and complex cross-shareholding relationships, which used to be a distinctive characteristic of Japan s stock market, have thus ebbed significantly. Third, insider ownership has surged from around 6% in 1999 to 11% in 2006. Until now, it was supposed that few of Japan s major firms were family-controlled and thus predisposed to both the positive and negative incentives mentioned earlier. However, our finding clearly indicates that they have grown in presence. Is this shift in ownership structure the result of a change in investor behavior? Are firms with widely dispersed ownership gradually transforming into family-controlled firms? Before jumping to any conclusions, we should check the transition in market composition of listed firms. Interestingly, the turnover rate of listed firms has been accelerating since 1999. Of the 1,233 sample firms in 1987, only 938 still existed in 2006 (295 firms, or 24% of the total, were delisted). Meanwhile, 643 new listings appeared in the same period. This heavy turnover is thought to significantly affect the ownership structure. In fact, with regard to the original 1,233 listed firms in 1987 (called traditional firms below), three trends are evident. First, the trend in institutional ownership does not significantly differ from what was described earlier. Second, while cross-shareholding has declined over the period, the decline is less pronounced, and halts at 11%. Thus at traditional firms, cross-shareholding has unwound by less than generally perceived. Moreover, recently it has actually rebounded from a low of 11.0% in 2004 NLI Research 3 2008.03.31

Exhibit 2 Trend in Ownership Structure of Public Companies (First Section) 25% 20% Institutional investor 15% Insider 10% 5% 0% 1987 1990 1993 1996 1999 2002 2005 Listed firm Financial institution Stock ownership plan Notes: Shows average shareholding ratios at listed firms on the First Section of Tokyo, Osaka, and Nagoya stock exchanges. Financial institutions exclude cross-shareholders. Listed firms exclude financial institutions and cross-shareholders. Fiscal year No. of firms Financial institution Listed firm Foreign firm Insider Stock ownership plan Crossshareholder Crosssharehold. Government Institutional investor (%) Other minority sharehold. 1987 1,233 14.54 6.56 8.18 0.66 6.59 1.21 0.11 5.85 56.30 1988 1,268 14.79 6.31 8.06 0.67 6.67 1.16 0.10 6.46 55.80 1989 1,307 14.70 6.18 8.11 0.61 6.82 1.10 0.09 7.73 54.66 1990 1,342 14.97 6.23 8.25 0.60 6.93 1.11 0.09 7.95 53.88 1991 1,381 14.84 6.22 8.45 0.56 7.10 1.20 0.09 8.64 52.90 1992 1,387 15.05 6.20 8.32 0.56 7.05 1.34 0.09 8.64 52.76 1993 1,301 15.03 6.27 8.11 0.58 5.96 1.31 0.09 10.08 52.55 1994 1,301 15.16 6.15 7.96 0.58 5.88 1.36 0.16 10.39 52.36 1995 1,322 14.89 5.80 8.20 0.58 6.03 1.45 0.15 10.85 52.06 1996 1,356 14.65 5.54 8.44 0.60 6.08 1.50 0.14 11.40 51.65 1997 1,393 14.29 5.26 8.56 0.65 6.47 1.69 0.14 10.96 51.97 1998 1,405 13.41 5.22 8.83 0.68 6.63 1.91 0.14 10.49 52.68 1999 1,459 12.56 4.79 8.99 0.72 7.48 2.00 0.13 11.33 51.99 2000 1,523 11.64 4.43 9.39 0.78 8.44 2.14 0.11 12.23 50.82 2001 1,549 11.21 4.20 9.32 0.86 9.14 2.21 0.12 13.08 49.87 2002 1,570 10.09 4.03 9.11 0.86 9.91 2.33 0.11 13.89 49.67 2003 1,594 9.75 3.35 9.01 0.71 10.31 2.31 0.14 16.26 48.16 2004 1,687 8.87 3.07 9.33 0.72 10.50 2.20 0.14 18.76 46.40 2005 1,734 8.70 2.84 9.16 0.57 10.80 2.03 0.12 21.27 44.50 2006 1,768 8.65 2.66 9.26 0.59 11.09 1.97 0.13 21.81 43.84 to 11.3% in 2006. Third, insider ownership at traditional firms fell from 6.59% in 1987 to 3.57% in 2006, led by a decline in ownership by directors. However, this trend does not diminish the importance of the growth of family-controlled firms. Their rising prominence, accompanied by the increasing number of newly listed firms, could unleash a new phase of conflict of interest between controlling and minority shareholders. 6 4. Ownership Structure and Firm Value We next apply a quantitative model to measure the impact of ownership trends on management performance (return on assets). 7 Our focus is on the decade from 1997 to 2006, when structural changes were most pronounced. Due to comparison problems with the financial data, financial industries are excluded from the analysis. In addition, firms listed on the First Section for ten years or less are excluded because NLI Research 4 2008.03.31

their typically high insider ownership and strong business performance would bias results. 8 investors all serve to enhance management discipline. In Exhibit 3, the bar graph shows the change of ROA when each shareholder ratio is increased by 1-percentage point. Pale bars indicate that results are statistically insignificant and are not sufficient evidence of shareholder effect. As the graph shows, three shareholder types have a statistically significant effect on ROA performance: cross-shareholders, financial institutions, and institutional investors. 9 The effect on ROA is negative when cross-shareholders and financial shareholders increase, but positive when institutional investors increase. For example, if the average cross-shareholding ratio of 15% is completely unwound, ROA would rise 0.25-percentage point (0.017 x 15). This amounts to a 7% improvement in the average ROA of 3.7%. A similar magnitude of ROA change is obtained with the other two shareholder types. Results suggest that unwinding of cross-shareholdings, reduction of long-term shareholdings of financial institutions, and increase in presence of institutional Next, we examine how the market evaluates changes in shareholder type. Using Tobin s Q ratio (total market value plus total liabilities, divided by total assets), we apply the same statistical model to estimate the impact on market value (Exhibit 4). 10 Results suggest that the market tends to discount firm value when insider influence is strong, and conversely, to apply a premium when outsider influence prevails. Insiders include cross-shareholders, parent firms, directors and their families, and shareholding by the employees or business partners of the firm. The market will discount firm value if minority shareholders are exposed to the risk of exploitation from insiders. 11 On the other hand, institutional investors and foreign firms are seen as outsiders. If the market believes they will monitor management effectively, it applies a premium valuation. Moreover, the magnitude of these shareholder types is not trivial. For example, unwinding the cross-shareholding ratio from 15% to zero would Exhibit 3 Effect of a 1-Percentage Point Shareholder Increase on ROA Change in ROA (%) 0.015 0.011 0.011 0.010 0.007 0.005 0.000 0.001-0.005-0.010-0.008-0.015-0.020-0.025-0.017-0.021-0.015 Cross- Financial Listed Foreign Insider Stock Govt. Institutional shareholder institution firm firm owner. plan investor Note: Pale bar indicates result is not statistically significant. NLI Research 5 2008.03.31

increase Tobin s Q by 0.072 (0.0048 x 15), the equivalent of an 8% increase in the average firm s market value (Q=0.92). In addition, as Exhibit 4 shows, greater institutional ownership would have a comparable impact. By comparison, an increase in stock ownership plans (shareholding by employees or business partners of the firm) not only has a negative effect, but that effect is three times greater than the positive effect of either of the above. However, considering that the average shareholding ratio of stock ownership plans is only about 2% and low in volatility, a jump of 10-percentage points is highly improbable. Indeed, the actual effect of stock ownership plans is thought to be small compared to that of cross-shareholders and institutional investors. The effect is even smaller for listed firms, foreign firms, and insiders, which all have statistically significant results. First, the surging presence of institutional investors has helped to discipline management and enhance firm value. Despite differences in the level of monitoring activity and influence, institutional investors share a common objective of improving the return on investment. If inefficient management becomes an issue, institutional investors might herd together (that is, independently decide to take the same action) to confront incumbent management. This constantly keeps management on its toes. Second, along with growing institutional ownership, the unwinding of cross-shareholdings has also served to enhance management performance. As stable shareholdings unwind, management becomes increasingly exposed to market monitoring. The recent resurgence of cross-shareholding at some firms is likely to hurt their management efficiency. 5. Conclusion By approaching the available data from a new perspective, we confirmed that corporate ownership has changed significantly in recent years. Below we briefly discuss the implications of our results. Third, family-controlled firms have become more conspicuous amid the growing number of new listings. They are controlled by directors and their families, or private firms controlled by individuals. This trend is particularly significant in light of the recent growth of stock markets that foster new venture businesses and small and medium enterprises. Exhibit 4 Effect of a 1-Percentage Point Shareholder Increase on Market Value Change in market value (%) 0.006 0.004 0.002 0.0004 0.000-0.002 0.0018 0.0018-0.0008-0.0008 0.0040-0.004-0.006-0.0048-0.008-0.010-0.012-0.014-0.016-0.0132 Cross- Financial Listed Foreign Insider Stock Govt. Institutional shareholder institution firm firm owner. plan investor Note: Pale bar indicates result is not statistically significant. NLI Research 6 2008.03.31

The latest research suggests that this phenomenon may call up the old yet new governance problem in which dominant shareholders take advantage of minority shareholders. Up to now, the governance framework has addressed the conflict of interest between shareholders and management. However, this framework has little bearing on the burgeoning conflict between dominant shareholders and minority shareholders. Family control may prove to be a threat to minority shareholders even here in Japan. Amid vast institutional changes that have appeared in the main bank system, lifetime employment system, and seniority wage system, is the protection of the interests of minority shareholders fully sufficient? Many thorny issues may still remain concealed in this regard. Our analysis marks the first step of a new approach to analyzing corporate ownership and its implications. Further research will be necessary to confirm and expand on our tentative conclusions. References Claessens, S., S. Djankov, J. P.H. Fan, and L. H.P. Lang. 2002. Disentangling the incentive and entrenchment effects of large shareholdings. Journal of Finance 57: 2741-2771. Claessens, S. and J. P.H. Fan.2002. Corporate governance in Asia: A survey, International Review of Finance 3: 71 103. Johnson, S., R. La Porta, F. Lopez-de-Silanes and A. Shleifer. 2000. Tunneling. American Economic Review 90: 22-27. La Porta, R., F. Lopez-de-Silanes and A. Shleifer. 1999. Corporate ownership around the world. Journal of Finance 54: 471-518. (In Japanese) Nitta, K. 2000. Cross-shareholding and corporate performance. Security Analyst Journal 38 (2): 72-93. Miyajima, H., and K. Nitta. 2003. Governance problems in low-performance industries. Security Analyst Journal 41 (12): 29-47. Miyajima, H., K. Nitta, N. Saito, and Y. Omi. 2004. Corporate governance and management efficiency: Effects and channels of corporate governance, and the effect of corporate characteristics. Nissay Kisokenkyusho Shoho 33: 52-98. Endnotes 1. There are some recent papers that discuss shareholder types. See La Porta et al. (1999), Claessens et al. (2002), and Claessens and Fan (2002). However, since they do not address features unique to Japanese firms, we do not use them here. 2. The minimum threshold is set at 3% because a lower value would significantly increase the cost of research. In addition, a 3% shareholding ratio conveys important rights to the shareholder including the right to inspect books, convene shareholder meetings, call for the dismissal of executives, claim for turnaround, and present shareholder proposals. 3. Cross-shareholding is defined as a state in which two public companies mutually own each other s shares. It is confirmed using our proprietary shareholder composition analysis database, which combines data from the major shareholders database (Toyo Keizai Inc.) and financial statements database (Nikkei NEEDS). 4. Aggregate holdings of foreigners, directors, pension trusts, investment trusts, and employee stock ownership plans were obtained from the database of major shareholders (Toyo Keizai Inc.). 5. Compared to cross-shareholding ratios announced in the Fiscal 2003 Survey of Cross-Shareholding (NLI Research Institute), our values are smaller for fiscal 1997 and earlier, and larger for later years. Although the same data is used, the discrepancy stems from differences in the method and scope of compilation. 6. For more on the exploitation of minority shareholders by controlling shareholders, see Johnson et al. 7. Considering the heterogeneity of firms, we performed a panel regression analysis. The explained variable is the change of ROA from the previous period. The explanatory variables are all of the shareholder types mentioned above, with ROA, log value of total assets, and dummy variable for fiscal year as control variables. In all regression results, we chose the fixed-effect model over the random-effect model based on the Hausman specification test. 8. We did not analyze the implications of the large number of new listings on major stock exchanges by firms with a different ownership structure from traditional firms. This matter remains to be pursued in further research. 9. This result is consistent with previous empirical research. For details, see Nitta (2000), Miyajima and Nitta (2003), and Miyajima et al. (2004). 10. Since the problem is one of differences between companies, we performed an ordinary least squares regression analysis. The explained variable is Tobin s Q. The explanatory variables are all of the shareholder types mentioned above, with ROA, growth rate of total assets in the past three years, log value of total assets, industry dummy variable, and fiscal year dummy variable as control variables. 11. Even if the controlling shareholder does not actually exploit other shareholders, as long as the market shares this concern, the firm s market value will be discounted accordingly. NLI Research 7 2008.03.31